This is a section of the Cops Canvass Report
. See also San Diego County Canvass Procedure
We collected data in 85 precincts by scanning the report tapes that are produced by the Diebold (Premier Election Solutions) scanners, and then manually typed the data into our web-based database.
You can take a look at a summary of the precinct data here: http://www.copswiki.org/Cops/AllPrecincts
Simply click on a precinct sequence number and you can view the precinct data as entered by our volunteers. Click on attached documents to view the scanned data for each precinct.
The data for each race is pulled from the precincts in our sample. See that data here: http://www.copswiki.org/Cops/AllRaces
. You can click on a specific race to see the contribution from the sampled precincts that included that race. Note that because our data did not include the missing 9,000 ballots or any VBM or provisional ballots, our totals do not match the canvass report totals.
In addition, we had a difficult time collecting the data in this fashion. See the Race for Proposition 8 (http://www.copswiki.org/Cops/Race670
). We really tried to get all the data in on this one correct since it was so controversial. We had to leave out several precincts because the data was corrupted. In the first case, we simply could not read the scanner tape, for example.
Why we gathered this data
Our intention in gathering all this data was to see if we could detect tampering with the VBM and provisional ballots, since we had no way to review those. If we saw any race that was remarkably different from the certified results (like Prop 8 "No" winning) then an argument could be made that the VBM and provisionals, which have no audit trail that we can inspect, must have been fiddled with to get the result to flip the other way.
We weren't able to attempt to prove that hypothesis, but there was so many other issues we found, that the work was still of great benefit to bring to the attention of the public. It is our hope that in future elections, we will be able to get an exhaustive Audit Log of the central tabulator because those entries should exactly match the data scanned, and we can reconstruct the results of the election from that log. Such a 100% reconstruction would eliminate any meddling in the outcome of these elections by compromised elections officials who have access to the central tabulator.